National Security Response 127 Harv. L. Rev. F. 75

Creaky Leviathan: A Comment on David Pozen’s Leaky Leviathan

Responding to David Pozen, The Leaky Leviathan: Why the Government Condemns and Condones Unlawful Disclosures of Information, 127 Harv. L. Rev. 512 (2013)

Response To:


Over the past decade David Pozen has published a series of essays on themes related to state secrecy. These essays exhibit remarkable analytical rigor and feature deeply original and thought-provoking conclusions. Leaky Leviathan, the third in this series, is the most ambitious yet, offering a novel analysis of the regulatory regime governing the disclosure of classified information. Briefly put, Leaky Leviathan hypothesizes that the frenzy-inducing disclosures of classified information that America witnesses on a periodic basis owe less to institutional failures and more to the “deliberate choices made by high-level officials”1 who secretly benefit from the “permissive enforcement” of legal prohibitions against such disclosures.2 Pozen argues that this regulatory regime deserves more praise than critics give it because, in a number of subtle ways, it actually benefits national security, government efficiency, and democratic transparency. Consequently, “[i]f unauthorized disclosures were ever to be systematically suppressed,” he writes, “it would jeopardize so much more, and so much less, than First Amendment principles.”3

This essay probes both the analysis and the conclusion offered by Leaky Leviathan. I argue below that though high-level officials may not have an interest in zealously enforcing every disclosure of classified information, most observers would agree that many truly unauthorized disclosures adversely affect the “policymaking capacity” and “discrete agendas” of elements of the executive branch.4 This outcome makes one doubt that permissive enforcement owes all that much to deliberate choice on the part of high-level officials (at least where truly unauthorized disclosures are concerned). On the contrary, it lends weight to the view that the executive branch has been unable and unwilling to close the sluice gates due to legal and technological constraints that are easily underestimated, and due to political constraints emerging from strong passions in American society that exert pressure on democratically elected leaders.

Furthermore, even if Leaky Leviathan is right that permissive enforcement owes much to deliberate choice on the part of high-level officials, it is not clear that we are in a position to evaluate the overall consequences of this regulatory arrangement. There are a number of confounding variables to contend with, not the least of which is the secretive nature of the intelligence business, which often prevents a full accounting of the harm caused by unauthorized disclosures of classified information. Thus, as far as normative assessments are concerned, the most we can do, I argue, is become clearer about whether particular outcomes produced by this regulatory arrangement are tolerable.

I. A Summary of the Argument

Leaky Leviathan begins with a crucial observation: Though leaking has been described as a “problem of major proportions,”5 the reality is that high-level officials in “the executive branch have never prioritized criminal, civil, or administrative enforcement against leakers.”6 In fact, the indictment rate in this domain, Pozen notes, is close to “zero.”7 This statistic implies that even the Obama Administration’s much-decried “war on whistleblowers”8 is more akin to “a special operation than a war.”9

What explains the “dramatic disconnect”10 between what officials say about leaks and what they do about them? How, in other words, to explain the “negligible enforcement rate”11 of leak laws? The conventional answer, Pozen observes, “emphasizes how difficult it is to identify and prosecute leakers.”12 But this “constraint-based account,”13 he argues, is unsatisfactory for a number of reasons. To begin with, records indicate that government agencies report “only a small fraction” of leaks to the Justice Department, which implies that they are not entirely keen to track down leakers.14 Moreover, when complaints are actually lodged, investigations prove difficult because the Justice Department typically eschews “its most potent investigative tool” — namely, subpoenas.15 And even when leakers are identified, the Justice Department refrains from prosecution, purportedly due to concerns about introducing classified evidence in court. Yet, it displays little interest in bringing cases under statutes with less demanding requirements or in reforming the Classified Information Procedures Act (CIPA).16

The fact that so much “enforcement capacity” appears to be underutilized leads Pozen to suspect that the current regulatory regime must have “a significant intentional component.”17 Put another way, since the guard dog that is the Justice Department does not bite trespassers, even though it is equipped with a large and fine set of teeth, Pozen infers that it does not want to bite. It does not want to bite, he contends, because it has figured out that less biting means more bones. Put more formally, Pozen’s explanation — which is developed with far greater care and sophistication than this brief summary can convey — is that we ought to account for the “benefits that a strategy of minimal legal enforcement can yield for executive branch principals and the executive as an institution.”18 These benefits, he nicely shows, can be quite significant: disclosures of classified information can facilitate the execution of policies, enhance the legitimacy of the executive by lowering monitoring costs for overseers, promote communications within government and between the political branches, and even assuage important constituencies outside government, especially the media and civil society organizations.19

The foregoing account raises an obvious question. Few would deny that the executive branch benefits from “plants” (or authorized disclosures) and may even have an interest in permitting “pleaks” (or partially authorized disclosures) for the reasons Pozen identifies. But what interest does the executive branch have in letting “true leaks” (or unauthorized disclosures) go unpunished? Pozen’s key contention: “Plants need to be watered with leaks.”20 What this memorable formulation means is that if the executive branch only punished true leaks, then observers would easily discern that unpunished disclosures of classified information must be plants or at least pleaks — a discovery that would greatly limit the utility of such calculated disclosures. Thus, in order to be successful, “[t]he practice of planting requires,” Pozen writes, “some amount of constructive ambiguity as to its prevalence and operation.”21

Pozen recognizes that this innovative claim — that the executive branch engages in permissive enforcement partly in order to make room for plants and pleaks — might trouble advocates of national security and democratic transparency who would like to see robust and fairer application of the laws, respectively. Pozen responds to these concerns by identifying potential upsides to permissive enforcement that have not been sufficiently appreciated. For instance, leaks can potentially bolster national security by informing foreign powers of the scope and scale of American national security arrangements. Equally, plants and pleaks may have “educative and deliberative value for members of the public,”22 as they permit citizens to begin probing operations that are formally unacknowledged. These benefits, Pozen suggests, counsel against radical reform, which would in any case run contrary to the interests of key elements of the executive branch whose interests are served by the prevailing regulatory regime.

II. Does The Government Condone “True Leaks”?

The need for brevity means that I cannot discuss here many deeply admirable aspects of the analysis in Leaky Leviathan. I will focus only on two claims that I see as most in need of further investigation. The first is the claim that the prevailing regulatory regime serves the interests of the executive branch. “[A]n accommodating approach to enforcement,” Pozen writes, “has in the aggregate supported, rather than subverted, the government’s general policymaking capacity as well as many different policymakers’ discrete agendas.”23 This has happened because most “leaks” are in fact plants and “pleaks.” These kinds of disclosures serve as a “critical policymaking and communications tool” that allows parts of the executive branch to communicate with each other as well as with Congress, the media, and even foreign powers.24 An example here is the recent disclosure of the United States’ covert drone program by “unnamed U.S. officials,” which has allowed the executive branch to shape national and international perceptions of the program “without incurring the . . . risks that official acknowledgement may entail.”25 “[V]igorous enforcement of the leak laws,” Pozen claims, would “cripple” the government’s ability to make disclosures of this kind.26

This claim raises three concerns. First, are plants really violations of leak laws? Presidents (and their agents) might argue that they have the right to disclose information as they see fit.27 If plants are not counted as violations of leak laws, especially when such disclosures come from high-level officials that have the authority to declassify, then the mystery of underenforcement motivating Leaky Leviathan is somewhat reduced.

Suppose we accept that most disclosures of classified information violate leak laws, because the extent of authorization is unclear or contested — assume, that is, that most leaks are actually pleaks and true leaks. Still, the question arises: What executive branch interest is served by refraining from prosecuting true leaks that harm national security and embarrass government agencies and the White House? As explained earlier, in Pozen’s view consistently punishing true leaks would, among other things, lessen the credibility of plants and pleaks, as observers would then be able to figure out that a disclosure that goes unpunished must be a plant or a pleak. But does refraining from punishing true leaks really help the executive disguise plants and pleaks all that much? Surely examining whose interests a disclosure serves can help observers discern its source and thereby ascertain whether it is a plant or a pleak. Ordinary citizens may lack the time or interest to play this parlor game, but astute political observers — within the media, Congress, and foreign intelligence agencies — do not. For example, Senator John McCain, Charles Krauthammer, and Michael Massing, among others, have offered good reasons to believe the covert drone program was, in all likelihood, leaked by the White House.28

Maybe this analysis is too simplistic. Perhaps some true leaks really must go unpunished in order to facilitate “planting” and “pleaking.” Maybe the Yemenis and other interested observers are still not entirely sure who was responsible for leaking the details of the drone program, and maybe this uncertainty genuinely serves the interests of decisionmakers. Still, why does the executive permit what appear to be seriously damaging unauthorized disclosures — such as those concerning the National Security Agency’s sources and methods — to be published when 18 U.S.C. § 798 outlaws such conduct? There are, Pozen shows, an intricate set of norms and sanctions that the high-level officials use to police plants and pleaks.29 But these mechanisms of “social control,” he admits, have little effect on lower-level officials typically responsible for true leaks.30 So why, then, does the executive branch not fire a warning shot across the bow of the reporters, editors, and publishers that enable such reckless behavior?

Might legal and technological constraints be part of the explanation? Pozen gives this idea some credence — acknowledging a range of exogenous and endogenous constraints — but gives it lesser weight than conventional accounts do.31 If high-level officials were genuinely committed to plugging leaks, he suggests, “[i]t would be possible to stop a much higher percentage of disclosures.”32 I am not so sure. For instance, if the Justice Department started routinely issuing subpoenas, unauthorized disclosures would soon start being channeled through publishers overseas (indeed, this was Edward Snowden’s justification for making his disclosures to the Guardian). Under the circumstances, it makes sense for the Justice Department to stay its hand in the hope that American media organizations will display some small measure of responsibility (as opposed to organizations like the Guardian that have no long-term stake in the American political system). Similarly, the Justice Department could systematically employ polygraphs to uncover leakers, but even leaving aside the negative effects on general employee morale, there is the more basic problem that researchers “remain skeptical about any conclusion wrung from a polygraph.”33

What about political constraints — could they also be part of the explanation? Pozen considers the possibility, but expresses the concern that the swirl of interests and constraints makes it difficult to infer when executive branch hesitation can really be attributed to specifically “political” constraints.34 This is a reasonable concern, I admit. But we do not need a finely worked out theory of political constraints to discern how powerful media and partisan interests can obstruct the enforcement of the law. Pozen arguably underplays the ground realities when he writes:

[I]t is not clear that most media outlets would reflexively oppose greater enforcement against government insiders who supply some journalists with juicy information, yet who also jeopardize the industry’s legal and moral standing. The media would be on especially weak ground opposing greater administrative enforcement. The puzzle of permissive neglect remains.35

In practice, it is difficult to identify instances where media organizations have paid a meaningful price — moral, professional, or financial — for publishing rash or malicious disclosures. As far as I can tell, the Fourth Estate sees any and every effort to enforce the law against journalists as a gross violation of the First Amendment.36 Administrative enforcement can encounter similar counterpressure. A crackdown on self-proclaimed whistleblowers can invite the attention of organizations like the American Civil Liberties Union and the Government Accountability Project, who pound away until the President vows to refrain from punishing insubordination. Pozen himself provides a recent example when he highlights Attorney General Eric Holder’s declaration that “he has no desire for leak prosecutions to be his legacy.”37 How can this development be explained without referencing the steady drum beat of negative publicity that has apparently persuaded the Obama Administration to tread cautiously even when pursuing reporters responsible for some truly reckless disclosures?

The power of media organizations and activists to demand the permissive enforcement of leak laws would not be so great if there were not a wider culture of distrust when it comes to state secrecy. Here we must take seriously the skepticism and paranoia about state power and the zeal for transparency that have taken root in American society over the past half century. Arguably, these beliefs and passions, which manifest in diffuse and unpredictable ways, and are therefore not easily tested, serve to limit what democratically elected leaders can do to maintain secrecy. This picture leads me back to an analogy that I used earlier — perhaps the reason the guard dog that is the Justice Department does not bite trespassers is not because it does not want to, but because it has learned the hard way that biting trespassers leads to a mighty kick from the very same person who put it there to guard against trespassers.

III. Can We Judge “Permissive Enforcement”?

A second concern about Leaky Leviathan regards the claim that “middle range” theory allows us to better appraise the normative worth of the existing regulatory regime.38 It makes little sense, Pozen writes, to make normative judgments about a regime of permissive enforcement by focusing either on particular cases or on broad constitutional principles.39 The scholarship that focuses on particular cases suffers from a “fallacy of composition” — it makes the mistake of assuming that since some disclosures have proven to be rash or malicious, permissive enforcement is completely flawed.40 The scholarship that focuses on broad constitutional principles, by contrast, provides little practical guidance; in reality “cases and statutes . . . have mattered less in this area . . . than the legal literature’s fixation on them would suggest.”41 A proper evaluation of the existing regulatory regime, Pozen argues, “cannot get off the ground without a fuller understanding”42 of “leaks’ consequences, inside and outside government.”43

This claim raises three concerns. Does a detailed, empirical account of how and why leaks occur really help evaluate whether permissive enforcement has had desirable consequences? Pozen seems to believe so: he states that “an accommodating approach to enforcement has in the aggregate supported, rather than subverted, the government’s general policymaking capacity as well as many different policymakers’ discrete agendas.”44 Pozen certainly identifies a number of valuable executive branch interests that are plausibly served by leaks — and by pleaks in particular. Does the evidence support the broader claim though? Consider the longer view. Few would deny that trust in government has declined since the middle of the twentieth century. Arguably, this decline has been abetted in no small measure by unauthorized disclosures, such as those relating to Vietnam, Iran-Contra, and Iraq, which have eroded the executive branch’s legitimacy. Under the circumstances, can we be sure that a regime of permissive enforcement has really benefited the executive branch?45

Suppose this objection is dismissed as overly pessimistic. Perhaps there has been no decline in trust or, more plausibly, this decline is better attributed to other factors. Consider another question then: is it even possible to have an overall sense of the effects of permissive enforcement? Separate from the problem of tabulating the costs and benefits is the question of whether we can even know what the costs and benefits are. Pozen lists a number of national security “silver linings” associated with leaks. For example, leaks can “puncture a false sense of comfort,”46 be “information eliciting,”47 and can promote “deterrence by denial.”48 All this may be true, but until we fully understand the costs to be entered on the other side of the ledger, we will be hard pressed to say much about the overall consequences of permissive enforcement. For example, Snowden’s disclosures may have inadvertently served to intimidate Al Qaeda, but they may have also allowed the Chinese to better firewall their plans for the South China Sea. Similarly, Chelsea Manning’s disclosures may have facilitated public debate on the nature and purposes of American diplomacy, but they also led to the drying up of information sources. How are outsiders to know, much less weigh, these costs, especially when the damage caused by true leaks are usually better left secret than exposed to the world?

Consider the same question from the perspective of citizens. Pozen identifies a number of constitutional and democratic interests that can be furthered by leaks (plants and pleaks included). For example, disclosures of classified information can further “public accountability,”49 enable “critical discourse,”50 and expose the “pluralism and competition” associated with decisionmaking.51 Again, these claims seem perfectly reasonable, but recall that some of the most crucial national decisions of the past century — to intervene in Vietnam and to invade Iraq, for instance — were fostered by selective disclosures of classified information.52 These disclosures have led to victory and defeat, glory and sorrow, riches and ruin — but who can say with clarity what they have done for American democracy?

Let me be clear. Leaky Leviathan contains no hasty judgments. Pozen is admirably careful and measured in his praise of permissive enforcement; his broader objective is to show that the varied causes and consequences associated with leaks mean that the prevailing “system of information control defies simple normative assessment.”53 This conclusion I fully endorse. But the question remains: even if it advances our comprehension, does “middle-range theory” improve our ability to judge the value or worth of the prevailing “system of information control”? For the reasons outlined above, I think the answer must be no, since the overall consequences of permissive enforcement are obscure and hard to ascertain. As far as this “disorderly” system is concerned, the most we can do from a normative perspective, I think, is become clearer about whether particular outcomes it produces are more or less acceptable. We can do this by ascertaining the moral thresholds that unauthorized disclosures must meet — those disclosures that meet the specified thresholds ought to be celebrated (for instance, when a disclosure reveals serious moral or legal wrongdoing), and those disclosures that fail to meet the specified thresholds ought to be criticized or even punished (for instance, when a disclosure reveals classified information that does not reveal serious wrongdoing). To proceed in this fashion is to move forward on a case-by-case basis, not knowing where exactly the journey leads, but evaluating nonetheless the ports of call along the way.

IV. Conclusion

If the arguments outlined above are convincing then there is reason to question some of Leaky Leviathan’s key claims. As far as true leaks are concerned, the permissive enforcement of leak laws appears less a product of “deliberate choice” than of legal and technological constraints and strong passions in American society — forces that combine to engender a Leviathan that is not only leaky but also creaky. Furthermore, since it is difficult, if not impossible, to fully account for the costs of unauthorized disclosures, it is not clear that empirical or theoretical analysis of the prevailing “system of information control” can shed clear light on whether this regulatory arrangement ought to be commended or not.

Still, even if these aspects of Leaky Leviathan are open to question, its broader achievement is not. Among other things, Leaky Leviathan explains — and very convincingly so — why pleaks exist, why they are occasionally punished, how such punishments are administered, and why pleaks are valuable. These are hardly minor contributions to the scholarly literature on state secrecy. They constitute the first full-blown theory of pleaks — the unstable, but all-important, currency of Washington, D.C. From both methodological and substantive perspectives, then, Leaky Leviathan is a groundbreaking piece of scholarship that deserves the highest praise and the accolades that will doubtless come its way. It will or at least ought to become a standard reference piece on the subject.

* Assistant Professor of Politics, Princeton University; author, Rahul Sagar, Secrets and Leaks: The Dilemma of State Secrecy (2013).

1. David E. Pozen, The Leaky Leviathan: Why the Government Condemns and Condones Unlawful Disclosures of Information, 127 Harv. L. Rev. 512, 518 (2013).

2. Id. at 626.

3. Id. at 635.

4. Id. at 518.

5. Id. at 514 (quoting Statement on the Protection of Classified National Security Council and Intelligence Information, 18 Pub. Papers 22, 22 (Jan. 12, 1982)) (internal quotation mark omitted).

6. Id. at 515.

7. Id. at 536.

8. Id. (quoting, e.g., Glenn Greenwald, Secrecy Creep, Salon (Aug. 14, 2012, 9:17 AM), (internal quotation marks omitted).

9. Id.

10. Id. at 515.

11. Id. at 544.

12. Id.

13. Id.

14. Id. at 555.

15. Id. at 549.

16. See id. at 551—52, 554, 556—57.

17. Id. at 545.

18. Id. at 559.

19. See id. at 559—86.

20. Id. at 565.

21. Id. at 562.

22. Id. at 624.

23. Id. at 518.

24. Id. at 559.

25. Id. at 561.

26. Id. at 562.

27. As Pozen notes, “no scholarship has considered the law governing ad hoc declassifications — a strange lacuna in the secrecy literature.” Id. at 566 n.273. Some support for the view that high-level officials have the right to declassify as they see fit can be found in Department of the Navy v. Egan, 484 U.S. 518 (1988), which noted that the protection of classified information must be committed to the broad discretion of the agency responsible, and this must include broad discretion to determine who may have access to it. Id. at 529. More generally, see Michael A. Fletcher, Experts: Tactic Would Be Legal but Unusual, Wash. Post, Apr. 7, 2006, at A8; David E. Sanger & David Johnston, Bush Ordered Declassification, Official Says, N.Y. Times, Apr. 10, 2006, at A14; and Ronald Kessler, Former Prosecutor: Leak Case Is ‘Not Prosecutable,’ Newsmax (June 13, 2012, 12:07 PM), For an interesting debate on the moral and legal status of plants and pleaks, compare Editorial, A Good Leak, Wash. Post, Apr. 9, 2006, at B6, with Editorial, A Bad Leak, N.Y. Times, Apr. 16, 2006, at C11.

28. See Scott Shane, Renewing a Debate over Secrecy, and Its Costs, N.Y. Times, June 7, 2012, at A1 (quoting Senator McCain as saying that “administration officials were ‘intentionally leaking information to enhance President Obama’s image as a tough guy for elections’”); Charles Krauthammer, Barack Obama: Drone Warrior, Wash. Post (May 31, 2012),; Michael Massing, When Watchdog Meets Lapdog, Colum. Journalism Rev. (June 4, 2012, 2:50 PM),

29. Pozen, supra note 1, at 586.

30. Id.

31. For the conventional view, see Comm’n on the Intelligence Capabilities of the U.S. Regarding Weapons of Mass Destruction, Report to the President of the United States (2005); Report of the Interdepartmental Group on Unauthorized Disclosures of Classified Information (1982), reprinted in Presidential Directive on the Use of Polygraphs and Prepublication Review: Hearings Before the Subcomm. on Civil & Constitutional Rights of the H. Comm. on the Judiciary, 98th Cong. 166—80 (1985); James B. Bruce, How Leaks of Classified Intelligence Help U.S. Adversaries: Implications for Laws and Secrecy, in Intelligence and the National Security Strategist 399 (Roger Z. George & Robert D. Kline eds., 2006); Representative Pete Hoekstra, Remarks to the Heritage Foundation, Secrets and Leaks: The Costs and Consequences for National Security (July 25, 2005).

32. Pozen, supra note 1, at 518.

33. The Truth About Lie Detectors (aka Polygraph Tests), Am. Psychol. Ass’n (Aug. 5, 2004), On technological constraints, see also Adam Liptak, A High Tech War on Leaks, N.Y. Times, Feb. 11, 2012, at SR5.

34. Pozen, supra note 1, at 545.

35. Id. (footnote omitted).

36. Consider a recent example. In May 2012 the Associated Press published a news story containing details about an Al-Qaeda bomb plot — in spite of being asked by the White House to withhold the story out of concern for compromising sources and methods. The Justice Department subsequently subpoenaed the Associated Press’ phone records. Rather than criticize the AP’s disclosure of classified information, which did not even pretend to expose any wrongdoing, the Newspaper Association of America issued the following statement: “Today we learned of the Justice Department’s unprecedented wholesale seizure of confidential telephone records from the Associated Press. These actions shock the American conscience and violate the critical freedom of the press protected by the U.S. Constitution and the Bill of Rights.” Charlie Savage & Leslie Kaufman, Phone Records of Journalists Seized by U.S., N.Y. Times, May 14, 2013, at A1 (internal quotation marks omitted); see also Letter from Gary B. Pruitt, President and CEO, Associated Press, to Attorney General Eric Holder (May 13, 2013), available at

37. Pozen, supra note 1, at 632 (quoting Scott Shane & Charlie Savage, Administration Took Accidental Path to Setting Record for Leak Cases, N.Y. Times, June 20, 2012, at A14) (internal quotation marks omitted).

38. Id. at 546.

39. Id. at 634.

40. Id. at 565.

41. Id. at 634.

42. Id.

43. Id. at 514.

44. Id. at 518 (emphasis added).

45. This is not to suggest that permissive enforcement has hurt the executive branch. Rather, the point is that changing the time period under consideration can make it difficult to say what the consequences of permissive enforcement are in the aggregate. A separate and more troubling point, discussed below, is that we may not even know what the relevant costs and benefits are.

46. Pozen, supra note 1, at 610.

47. Id. at 611.

48. Id. at 612 (quoting Glenn H. Snyder, Deterrence by Denial and Punishment (1959)) (internal quotation marks omitted).

49. Id. at 624.

50. Id.

51. Id. at 621.

52. On Vietnam see Eric Alterman, When Presidents Lie 160—237 (2004); Daniel Ellsberg, Lying about Vietnam, N.Y. Times, June 29, 2001, at A23. On Iraq see Russ Hoyle, Going to War 265—89 (2008); Michael Isikoff & David Corn, Hubris: The Inside Story of Spin, Scandal, and the Selling of the Iraq War (2006).

53. Pozen, supra note 1, at 625.